The role of the pre-commissural fornix in episodic autobiographical memory and simulation

Abstract

Neuropsychological and functional magnetic resonance imaging evidence suggests that the ability to vividly remember our personal past, and imagine future scenarios, involves two closely connected regions: the hippocampus and ventromedial prefrontal cortex (vmPFC). Despite evidence of a direct anatomical connection from hippocampus to vmPFC, it is unknown whether hippocampal-vmPFC structural connectivity supports both past- and future-oriented episodic thinking. To address this, we applied a novel deterministic tractography protocol to diffusion-weighted magnetic resonance imaging (dMRI) data from a group of healthy young adult humans who undertook an adapted past-future autobiographical interview (portions of this data were published in Hodgetts et al., 2017a). This tractography protocol enabled distinct subdivisions of the fornix, detected previously in axonal tracer studies, to be reconstructed in vivo, namely the pre-commissural (connecting the hippocampus to vmPFC) and post-commissural (linking the hippocampus and medial diencephalon) fornix. As predicted, we found that inter-individual differences in pre-commissural - but not post-commissural - fornix microstructure (fractional anisotropy) were significantly correlated with the episodic richness of both past and future autobiographical narratives. Notably, these results held when controlling for non-episodic narrative content, verbal fluency, and grey matter volumes of the hippocampus and vmPFC. This study provides novel evidence that reconstructing events from one's personal past, and constructing possible future events, involves a distinct, structurally-instantiated hippocampal-vmPFC pathway.

Keywords: Episodic memory; Future thinking; Hippocampus; Mental time travel; White matter tractography; vmPFC.

Fig. 1

Examples of internal (episodic) and external details from past (A) and future (B) autobiographical narratives. The main event was required to be specific in time and place. If more than one specific event was provided, the event described in the most detail was coded as ‘internal’ and the other as ‘external’ (see Scoring). The main event (i.e., the event described in the most detail) is labelled for each example (e.g., ‘Sister's wedding’ for Past Example 1) and we underline the transcript to show where this event begins. The reader is referred to Levine et al. (2002, Table 1) and Hodgetts et al. (2017a, Table 1), for further details on scoring of subcomponent coding categories (subcomponent categories indicated here by bold text in brackets).

Fig. 2

Schematic illustration of the anatomical landmarks for fornix tract subdivision, and the connecting areas of interest. vmPFC = Ventromedial Prefrontal Cortex; MB = Mammillary Bodies.

Fig. 3

Waypoint region-of-interest (ROI) gates used for reconstructing the pre- and post-commissural fornix tract segments (Blue = SEED, Red = NOT, Green = AND). (For interpretation of the references to colour in this figure legend, the reader is referred to the Web version of this article.)

Fig. 4

Example reconstructions for the pre- and post-commissural fornix segments (Blue = Pre, Yellow = Post). (For interpretation of the references to colour in this figure legend, the reader is referred to the Web version of this article.)

Fig. 5

(A, B). Scatterplots depicting correlations between the number of details produced for the past versus the future AI narratives (A. Episodic, B. External) (N = 27). Marginal density is displayed on the opposite axis. Grey shading equals the 95% CI.

Fig. 6

(A–D). Scatterplots depicting the correlations of episodic past (A, B) and future (C, D) AI details with pre-/post-commissural fornix microstructure (fractional anisotropy, FA). Number of episodic past/future details (summed over 10 cue words) is plotted on the y-axis (N = 27). Grey shading equals the 95% CI.